Why your mithai tastes different every batch (and how to stop it)
Batch-to-batch variation is almost never a recipe problem. It is a measurement problem. Here are the six variables that actually move taste, and how to lock each one down.
Ask ten karigars how much khoya goes into a batch and you will get ten answers — all of them roughly correct, none of them the same. That gap is your inconsistency.
Taste variation comes from six places
1. Raw material variation
Khoya from two suppliers is not the same khoya. Fat content, moisture and freshness all differ, and they change your final product before you have done anything wrong.
Fix: written incoming standards for every major input, with a simple acceptance test at the gate. Track supplier-wise rejection month on month.
2. Unmeasured quantities
"Andaaz se" works brilliantly for one skilled person and fails completely across a team of twenty.
Fix: weigh everything. Every ingredient, every batch, recorded on the batch card.
3. Temperature
Sugar cook stage, roasting temperature, khoya heating — a ten degree difference changes texture, colour and shelf-life.
Fix: a thermometer per station and a target range written on the batch card. Not "till it looks right."
4. Time
How long the besan is roasted. How long the mixture rests. Time is free to control and almost nobody controls it.
Fix: timers, and a recorded start/end on the batch record.
5. Batch size drift
A recipe standardised for 20 kg does not scale linearly to 35 kg. Heat transfer changes, mixing changes, cook times change.
Fix: standardise at fixed batch sizes. If you need more output, run more batches — do not enlarge them informally.
6. Operator technique
The stage at which the mixture is pulled off heat, how it is spread, how it is cut.
Fix: demonstrate the standard once, then verify against it for two weeks. Technique sticks only when it is checked.
The one-page fix
A batch card. One sheet per batch, filled by the operator, signed by the supervisor:
| Field | Recorded |
|---|---|
| Batch no. & date | |
| Each input, weighed | |
| Cook temperature & time | |
| Yield (kg out) | |
| Sensory check — pass/fail | |
| Operator & supervisor |
Within a month you will have data. Within two, you will know exactly which shift, which supplier and which operator your variation is coming from.
Facing this in your factory?
Talk it through with Arjun Singh Gehlot. First conversation is free.